Date: Sun, 29 Jun 2003 20:24:42 -0400
From: Kurt Wall <kwall@kurtwerks.com>
To: Jeff Mock <jeff-ml@mock.com>
Cc: linux-kernel@vger.kernel.org
Subject: Re: PROBLEM: 2.4.21 ICH5 SATA related hang during boot
Message-ID: <20030630002442.GD276@kurtwerks.com>
References: <5.1.0.14.2.20030629135412.03c1d940@mail.mock.com> <5.1.0.14.2.20030629135412.03c1d940@mail.mock.com> <5.1.0.14.2.20030629165324.03da7cf0@mail.mock.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <5.1.0.14.2.20030629165324.03da7cf0@mail.mock.com>
User-Agent: Mutt/1.4i
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1920
Lines: 55

Quoth Jeff Mock:
> At 07:22 PM 6/29/2003 -0400, Kurt Wall wrote:
> >Quoth Jeff Mock:
> >>
> >> I'm running a 2.4.21 kernel on a redhat 9.0 system.
> >>
> >> I'm having a problem when using serial ATA drives on an Intel 875P/ICH5
> >> motherboard where the kernel will hang at approximately the same place
> >> in the boot process about 25% of the time.
> >
> >[tale of woe elided]
> >
> >[lots of snippage]
> >
> >> 1: nvidia: loading NVIDIA Linux x86 nvidia.o Kernel Module  1.0-4363  Sat
> >> Apr 19 17:46:46 PDT 2003
> >
> >You won't get a lot of help here until you lose this module.
> >
> >
> >> Linux agpgart interface v0.99 (c) Jeff Hartmann
> >> agpgart: Maximum main memory to use for agp memory: 1919M
> >> agpgart: Unsupported Intel chipset (device id: 2578), you might want to 
> >try
> >> agp_try_unsupported=1.
> >> agpgart: no supported devices found.
> >> 1: NVRM: AGPGART: unable to retrieve symbol table
> >
> >Hmm.
> 
> Guilty. The sad thing is that's just the tip of my politically
> incorrect iceberg.

It's not about politically incorrect. I could care less.

> The nvidia driver (and the attempt at agpgart) is loaded when X
> starts, long after the potential SATA related crash.  I changed my
> default init level and rebooted a few times to verify the crash, so
> I'm pretty sure that neither agpgart or the proprietary graphics driver
> are involved in the problem.

Except that without the driver source code, no one can be certain
that the NVIDIA modules isn't scribbling inappropriately in memory
and your particular setup just tickles an otherwise latent bug.

What's lost by *not* loading the driver?

Kurt
-- 
Does the name Pavlov ring a bell?
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/